Search

Data-driven thalamic nuclei parcellation depends on high-quality manual annotations. However, the small size and low contrast changes among thalamic nuclei, yield annotations that are often incomplete, noisy, or ambiguously labelled. To train a robust thalamic nuclei parcellation model with noisy annotations, we propose a label propagation algorithm based on random walker to refine the annotations before model training. A two-step model was trained to generate first the whole thalamus and then the nuclei masks. We conducted experiments on a mild traumatic brain injury~(mTBI) dataset with noisy thalamic nuclei annotations. Our model outperforms current state-of-the-art thalamic nuclei parcellations by a clear margin. We believe our method can also facilitate the training of other parcellation models with noisy labels.

Importance: Ultra-widefield fundus photography (UWF-FP) has shown utility in sickle cell retinopathy screening; however, image artifact may diminish quality and gradeability of images. Objective: To create an automated algorithm for UWF-FP artifact classification. Design: A neural network based automated artifact detection algorithm was designed to identify commonly encountered UWF-FP artifacts in a cross section of patient UWF-FP. A pre-trained ResNet-50 neural network was trained on a subset of the images and the classification accuracy, sensitivity, and specificity were quantified on the hold out test set. Setting: The study is based on patients from a tertiary care hospital site. Participants: There were 243 UWF-FP acquired from patients with sickle cell disease (SCD), and artifact labelling in the following categories was performed: Eyelash Present, Lower Eyelid Obstructing, Upper Eyelid Obstructing, Image Too Dark, Dark Artifact, and Image Not Centered. Results: Overall, the accuracy for each class was Eyelash Present at 83.7%, Lower Eyelid Obstructing at 83.7%, Upper Eyelid Obstructing at 98.0%, Image Too Dark at 77.6%, Dark Artifact at 93.9%, and Image Not Centered at 91.8%. Conclusions and Relevance: This automated algorithm shows promise in identifying common imaging artifacts on a subset of Optos UWF-FP in SCD patients. Further refinement is ongoing with the goal of improving efficiency of tele-retinal screening in sickle cell retinopathy (SCR) by providing a photographer real-time feedback as to the types of artifacts present, and the need for image re-acquisition. This algorithm also may have potential future applicability in other retinal diseases by improving quality and efficiency of image acquisition of UWF-FP.

Normal pressure hydrocephalus~(NPH) is a brain disorder associated with enlarged ventricles and multiple cognitive and motor symptoms. The degree of ventricular enlargement can be measured using magnetic resonance images~(MRIs) and characterized quantitatively using the Evan's ratio (ER). Automatic computation of ER is desired to avoid the extra time and variations associated with manual measurements on MRI. Because shunt surgery is often used to treat NPH, it is necessary that this process be robust to image artifacts caused by the shunt and related implants. In this paper, we propose a 3D regions-of-interest aware (ROI-aware) network for segmenting the ventricles. The method achieves state-of-the-art performance on both pre-surgery MRIs and post-surgery MRIs with artifacts. Based on our segmentation results, we also describe an automated approach to compute ER from these results. Experimental results on multiple datasets demonstrate the potential of the proposed method to assist clinicians in the diagnosis and management of NPH.

The accurate prediction of vehicle state based on the data acquired by the vehicular networking system plays an important role in improving traffic safety in the transportation section. However, it is difficult to accurately predict the vehicle state due to the highly dynamic road environment and various drivers’ behaviors. To this end, in this article, we propose a two-level quantized adaptive Kalman filter (KF) algorithm based on the autoregressive moving average (MA) model to predict the vehicle state (including the moving direction, driving lane, vehicle speed, and acceleration). First, we propose a vehicular networking system to acquire the vehicle data by exchanging traffic data between the onboard unit and the roadside unit (RSU). Then, we predict the vehicle state at the edge cloud server (ECS) equipped at the RSU. Specifically, we utilize the autoregressive MA model to predict vehicle acceleration at the next moment. Then, the predicted vehicle acceleration is used as an input variable of the adaptive KF model to predict the vehicle location and speed at the next moment, in which we quantify the predicted vehicle location to the moving direction and the driving lane. Finally, the ECS broadcasts the predicted state to other RSUs. Through the communication with the road unit, all vehicles moving at the intersection can share vehicles states each other. In this doing, we can efficiently improve traffic safety in the intersection. We provide numerical simulations to validate the effectiveness of the autoregressive MA model used for predicting acceleration. Then, we evaluate the efficiency of the proposed two-level quantized adaptive KF algorithm. Compared with five conventional prediction algorithms, our proposed algorithm can improve the speed prediction accuracy by 90.62%, 89.81%, 88.91%, 82.76%, and 70.77%, respectively, which implies that our algorithm is a promising scheme for predicting the vehicle state in vehicular networks.

Nonorthogonal multiple access (NOMA) and mobile edge computing (MEC) have been emerging as promising techniques in narrowband Internet of Things (NB-IoT) systems to provide ubiquitously connected IoT devices with efficient transmission and computation. However, the successive interference cancellation (SIC) ordering of NOMA has become the bottleneck limiting the performance improvement for the uplink transmission, which is the dominant traffic flow of NB-IoT communications. Also, in order to guarantee the fairness of task execution latency across NB-IoT devices, the computation resource of MEC units has to be fairly allocated to tasks from IoT devices according to the task size. For these reasons, we investigate the joint optimization of SIC ordering and computation resource allocation in this paper. Specifically, we formulate a combinatorial optimization problem with the objective to minimize the maximum task execution latency required per task bit across NB-IoT devices under the limitation of computation resource. We prove the NP-hardness of this joint optimization problem. To tackle this challenging problem, we first propose an optimal algorithm to obtain the optimal SIC ordering and computation resource allocation in two stages: the convex computation resource allocation optimization followed by the combinatorial SIC ordering optimization. To reduce the computational complexity, we design an efficient heuristic algorithm for the SIC ordering optimization. As a good feature, the proposed low-complexity algorithm suffers a negligible performance degradation in comparison with the optimal algorithm. Simulation results demonstrate the benefits of NOMA in reducing the task execution latency.

Real-time and accurate short-term traffic prediction can effectively improve traffic efficiency, reduce accidents, and facilitate relevant departments to take reasonable traffic guidance measures. Therefore, we propose a two-level data driven model for short-term traffic prediction in an edge computing environment. Firstly, a Deep Belief Network (DBN) is developed to extract the traffic characteristics between the road occupancy and road flow collected by the deployed detectors. Then, we predict the developed future road flow of each road segment based on the output of the DBN, which would be used as one of the inputs of a Hidden Markov Model (HMM). Finally, a HMM is developed to predict the future road speed of each road segment characterizing the statistical relationship between the road flow and road speed. To validate the effectiveness of our proposed model, the data from the Performance Measurement System (PeMS) of the California Department of Transportation is applied. Simulation results show that our proposed model has better prediction performance in short-term traffic prediction than other models.

This paper studies mobile edge computing (MEC) networks where multiple wireless devices (WDs) choose to offload their computation tasks to an edge server. To conserve energy and maintain quality of service for WDs, the optimization of joint offloading decision and bandwidth allocation is formulated as a mixed integer programming problem. However, the problem is computationally limited by the curse of dimensionality, which cannot be solved by general optimization tools in an effective and efficient way, especially for large-scale WDs. In this paper, we propose a distributed deep learning-based offloading (DDLO) algorithm for MEC networks, where multiple parallel DNNs are used to generate offloading decisions. We adopt a shared replay memory to store newly generated offloading decisions which are further to train and improve all DNNs. Extensive numerical results show that the proposed DDLO algorithm can generate near-optimal offloading decisions in less than one second.

Powering relays with harvested renewable ambient energy has been emerging as a promising solution to reduce the on-grid energy consumption and greenhouse gas emissions in green relaying communication networks. In this paper, we study the joint time scheduling and power allocation problem for the Decode-and-Forward energy-harvesting relay communication network. Particularly, our goal is to maximize the end-to-end throughput by a deadline subject to the finite data and energy storage. Due to the multi-slot optimization, the traditional deep reinforcement learning (RL) framework cannot be directly applied to obtain the optimal solution of maximizing the end-to-end throughput by a deadline in the online manner. To this end, we explore a novel deep reinforcement learning framework consisting of multiple computation units to obtain the online time scheduling and power allocation based on the current causal knowledge of energy arrivals and channel fading at each time slot. Simulation results show that the proposed deep reinforcement learning based algorithm can achieve more than 90% of maximum end-to-end throughput.

Worldwide, 80% of patients who undergo surgery receive opioid analgesics as the fundamental agent for pain relief. However, the irrational use of opioids leads to excessive drug dependence and drug abuse, resulting in an increased mortality rate and huge economic loss. The crisis of opioid overuse remains a great challenge. In this review, we summarize several key factors in opioid abuse, including race, region, income, genetic factors, age and gender, smoking and alcohol abuse, history of chronic pain and analgesic drug abuse, surgery, neuropsychiatric illness, depression and antidepressant use, human factors, national policies, hospital regulations, and health insurance under treatment of pain. Furthermore, we present several prevention strategies, such as perioperative measures, opioid substitutes, treatment of the primary illness, emotional regulation, use of opioid antagonists, efforts of the state, hospitals, doctors and pharmacy benefit managers, gene therapy, and vaccines. Greater understanding and better assessment are required of the risks associated with opioid abuse to ensure the safety and analgesic effects of pain treatment after surgery.

This paper focuses on the design of radio frequency identification (RFID) data-driven vehicle speed prediction method using adaptive Kalman filtering. First of all, when the vehicle moves through a RFID tag, the reader needs to acquire the state information (i.e., current speed and time stamp) of the last vehicle across the tag, and meanwhile transmits its state information to this tag. Then, the state space model can be formulated according to the acquired state information. Finally, the adaptive Kalman filtering algorithm is proposed to predict and adjust the speed of vehicles. Adaptive Kalman filtering algorithm achieves the adaptive updating of variable forgetting factor by analyzing the error between the expected output value and the actual output value, so as to achieve the online updating of the prediction model. The numerical results further show that compared with the conventional Kalman filtering algorithm, the proposed algorithm can increase the speed prediction accuracy by 20%. This implies that the proposed algorithm can provide the better real-time effectiveness for the practical implementation.

The traditional speed prediction generally utilizes GPS and video images, and thus the prediction accuracy is heavily dependent on environmental factors. To this end, through using RFID (Radio Frequency Identification) data, this paper proposes a vehicle speed prediction algorithm based on Extended Kalman Filter (EKF). Specifically, the proposed algorithm works as follows. First, the RFID reader equipped in the vehicle acquires the state information of tags deployed on the road. Second, The data processing module equipped in the vehicle demodulation and decoding these information. At the same time, the RFID reader sends information to the RFID label after the current information is encoded and modulated. Third, the vehicle predicts the vehicle speed based on the EKF through establishing the state space model with acquired state data. The simulation results show that the proposed algorithm can effectively predict the vehicle speed at 0.6 s.

The traditional speed prediction generally utilizes the Global Position System (GPS) and video images, and thus, the prediction precision mainly depends on environmental factors (i.e., weather, ionosphere, troposphere, air, and electromagnetic waves). We study the Radio Frequency Identification (RFID) data-driven vehicle speed prediction and proposed an improved extended kalman filter (i.e., the adaptive extended kalman filter, AEKF) algorithm. Firstly, the on-board RFID reader equipped in the vehicle reads the information (i.e., current speed and time) from the tag deployed on the road. Secondly, the received information is transmitted to the on-board information processing unit, and it is demodulated and decoded into available information. Finally, based on the vehicle state space model, the AEKF algorithm is proposed to predict vehicle speed and improve the prediction results, so that the vehicle speed gradually approaches the actual vehicle speed. The simulation results show that compared with the conventional extended kalman filter (EKF) algorithm, our proposed AEKF algorithm improves the dynamic performance of the filtering and better suppresses the filtering divergence process. Moreover, the AEKF algorithm also improves the precision of the Mean Square Error (MSE) and Mean Absolute Error (MAE) by 57.4% and 32.4%, respectively.